One of the story lines emerging in the last couple of days about the midterm elections is the possibility that a lot of races are going to be pretty close. So for example, “Politico reports”:http://www.politico.com/news/stories/1010/44231.html:
bq. …it’s become more apparent that just a few percentage points could end up swinging the outcome in many races.
In a closely related vein, “Nate Silver notes”:http://fivethirtyeight.blogs.nytimes.com/2010/10/26/tonights-house-forecast-52-seat-gain-for-g-o-p/
bq. Since there are a very large number of competitive seats, relatively small anomalies in the polling could potentially affect the outcome of dozens of races. Although the Democrats’ overall position is poor, it is not yet so poor that it couldn’t be salvaged if they beat their polling averages by 2 or 3 points nationwide.
Silver’s comment about small anomalies in polling reminded me of a “Pew Research Center report”:http://pewresearch.org/pubs/1761/cell-phones-and-election-polls-2010-midterm-elections I read earlier this month about the potential pitfalls of polling that doesn’t take account of cell phone only users. Here’s the data from the Pew study:
One survey shows no difference between the land line only and combined polls. The other three, however, shows a difference of anywhere from 4-6 percentage points, all favoring the Republican party. Moreover, the one set of data they released for likely voters shows this effect is not limited to registered voters.
If 2008 witnessed the rise of 538.com and attempts to leverage massive amounts of polling data to estimate control of the electoral college, then 2010 likely represents our first real attempt to apply these types of methods in an off-year congressional election. Moreover, cell phone only households have undoubtedly grown since 2008; Pew reports that today a quarter of US households “can not be reached by a land line”:http://voices.washingtonpost.com/behind-the-numbers/2010/10/pew_cellphone_bias_may_be_bigg.html/. Taken together, if there are a lot of seats within 2-3 percentage points in the polls and it turns out that the pollsters have not adequately corrected their data for the types of effects cell phone only effects identified by Pew, then if – and this is obviously a big if – Democratic candidates do better than expected on Nov. 2nd, polling issues related to cell phone only households might be one possible explanation.
